This application proposes to design and implement a series of simulation studies and secondary data analyses to compare different design and analysis approaches to sequential decision making in the context of alcohol treatment. In this proposal, we focus on selection of treatment regimes, using Bayesian and causal models. Our goal is to demonstrate that these approaches can be understood and implemented by clinical researchers.